World Cup Soccer home advantage.

Brown, Terry D., Jr. ; Van Raalte, Judy L. ; Brewer, Britton W. 等

Home advantage (HA) in sport competition has been a well-documented phenomenon (Carron & Hausenblas, 1998). In an informal content analysis of media reports, Edwards and Archambault (1989) found that more references were made to the difficulty of defeating a team on their home ground than any other single factor, including talent, prior record, injuries, and momentum.

Koppet provided one of the first operational definitions of HA in 1972, "being at home increases your chance of winning" (p. 1C). Courneya and Carron (1992) refined the definition of HA with the following, "home teams in sport competitions win over 50% of the games played under a balanced home and away schedule" (p. 13). Researchers using this definition have found a HA in many different sports and at many levels of competition (Courneya & Carron, 1992). Because some teams (including World Cup soccer teams) do not play a balanced schedule, definitions of HA may have to be modified to include these teams. Also, other questions about the Courneya and Carron definition of HA have been raised (Bray, 1999).

Bray (1999) noted that the Courneya and Carron (1992) definition of HA was based on the home win percentages (HWP) of entire leagues over a number of years. Individual team data (overall and by year) were overlooked and no account for the away record was considered. To remedy these problems, Bray defined HA as when the HWP minus the away win percentage (AWP) is greater than 5%. By using this equation, the away record is taken into account, and individual team statistics can be calculated overall and by year.

Regardless of the definition of HA used, teams competing at home have an advantage over teams competing away from home (Bray, 1999; Courneya & Carron, 1992; Koppet, 1972). Several factors have been hypothesized to explain this HA effect, including familiarity with the playing facility (Moore & Brylinski, 1995), game importance (Baumeister, 1995), and travel fatigue (Pace & Carron, 1992).

The purpose of this research was to explore several factors related to HA with data collected from World Cup soccer teams. World Cup soccer teams are of particular interest because they have varied schedules, competing in their home countries, away, and in neutral countries. Comparisons can be made among these three settings. Second, World Cup teams play games that vary in importance from a non-championship game (friendly or qualifier) to a championship game (Continental championship or World Cup Championship). Relationships between game importance and HA can be explored. Finally, because World Cup soccer teams travel long distances, the effects of travel on HA can be examined.

Method

Database

The data for this study were obtained from the official Federation Internationale de Football Association (FIFA) Coca-Cola Rankings internet site (FIFA, 1999). The data included the results of 3,914 games played by 32 international soccer teams between January 1987 and the end of the 1998 World Cup in France (July 12, 1998). The 32 teams included were the teams that participated in the 1998 World Cup in France: Argentina, Austria, Belgium, Brazil, Bulgaria, Cameroon, Chile, Colombia, Croatia, Denmark, England, France, Germany, Iran, Italy, Jamaica, Japan, Mexico, Morocco, Netherlands, Nigeria, Norway, Paraguay, Romania, Saudi Arabia, Scotland, South Africa, South Korea, Spain, Tunisia, USA, and Yugoslavia. The variables included in this study were: (a) outcome for each game (win, lose, draw), (b) location of the game (home, away, or on a neutral site), (c) the city in which the game occurred, (d) number of goals scored by and scored upon each team, (e) date of the game, and (f) type of game, as classified by FIFA.

Data Collection and Procedures

The data were printed from an internet archive site (FIFA, 1999) and entered into a database for all 32 teams for the years specified. Nonofficial FIFA games and Olympic games were omitted because these games are not recognized by FIFA in ranking national soccer teams (FIFA, 1999) and because the players often differ from the players that compete in official FIFA matches.

To assess distance traveled, the great circle distance between cities where sequential games were played was calculated. According to Weisstein (1999), the great circle distance is the shortest surface arc between two points on a sphere. For this study, the distance between the previous game and current game was used if the previous game was within 15 days. After 15 days the distance from the home stadium was used, because it was assumed that if the next match was over 15 days away, the team would go home before leaving for that match.

Once these calculations were made, all data were collapsed by team. Then, HWP and AWP were calculated for every team. From these calculations, HWP minus AWP (HWP - AWP) was calculated. Using criteria from Brown et al. (1999), each team was categorized as having a HA, neither advantage nor disadvantage (NAD), or home disadvantage (HD). Finally, the data were analyzed using non-parametric statistics due to the positive skew of the data (see Table 1).

Results

Descriptive statistics on the distance (km) that teams had traveled from their previous game, the number of days since their previous game, the number of goals scored against them, the number of goals scored for them, and the goal differential are presented in Table 1. On average, teams had a little over a month between games, and traveled over 1700 km to get to games. The 32 World Cup teams analyzed scored more goals than they had goals scored against them.

The number of games played at home, away, and neutral sites and the outcomes of those games are presented in Table 2. The 32 World Cup teams analyzed played more games at home than they did away or on neutral sites. Also, these teams won more games at home than they did at away or on neutral sites.

Overall, teams won 48% of all their games. They won 63% at home, 37% away, and 40% at neutral sites. In terms of HWP - AWP, teams had a mean difference of 27%. Based on this difference, teams were classified into HA categories (HA, NAD, HD). All teams, except one, were classified as having a HA. The other team was classified as having NAD.

Familiarity with the Playing Facility

To test for familiarity with facilities, a 3 (location of game: home, away, neutral) x 3 (outcome: win, lose, draw) chi-square test was performed. The association between the two variables was significant ([chi square](4, n = 3914) = 206.90, p<.001, C = .23). The number of games won at home exceeded the expected count. At neutral sites, teams won about the same number of games as the expected count. At away sites, the number of wins was much less than the expected count. In terms of drawing a game, all teams had about the same frequency as the expected count at all three sites. Thus, teams won more and lost fewer games at home, and lost more and won fewer games away.

Game Importance

Chi-square tests were also conducted to examine the relationships between game importance (championship, nonchampionship) and outcome (win, lose, draw) for games played at home, away, and at neutral sites. No association was found between importance of game and outcome at home ([chi square](2, n = 1490) = 0.75,p>.05) or at neutral sites, ([chi square] (2, n = 1000) = 5.86,p>.05). For games played away, however, a significant association was found ([chi square] (2, n = 1342) = 6.50, p < .04, V = .07). Away teams lost more and won fewer championship games than the expected count. In non-championship games, away teams won, lost, and tied about the same number of games as the expected count.

Travel

To analyze the travel factors data, Pearson r correlations were calculated among goals against, goals for, goal differential, distance traveled from the previous game if within 15 days, or from home if over 15 days, and number of days since the previous game. As seen in Table 3, distance wad significantly correlated with goals against (r = .09, p < .001), goals for (r = -.08, p <.001), and goal differential (r = -.11, p <.001). Number of days since the previous game was significantly correlated with goals against (r = -.04, p = .02).

Additional analyses of travel factors were completed by using Kruskal-Wallis one-way ANOVA by ranks tests because the data were too skewed to use parametric statistics (see Table 4). In the first Kruskal-Wallis test, distance from the previous city or home was compared across the levels of outcome (win, lose, or draw) in all games. Significant differences were found (H(2) = 99.36, p < .001, [[eta].sup.2] = .03). Results indicated that teams traveled shorter distances for games that they won than for either games that they tied or lost. No difference in distance traveled was found between tied and lost games.

For the second Kruskal-Wallis test, number of days from the previous game was assessed for games outcome (won, lost, and tied) in all games. Significant differences were found here as well (H(2) = 6.97, p < .04, [[eta].sup.2] = .002). Upon further analysis, it was found that more days had elapsed between games prior to won games than prior to games that were tied or lost. No difference in days between games was found between tied and lost games.

Discussion

The study was designed to examine two aspects of World Cup soccer HA. First, World Cup soccer teams were examined to determine if they have a HA. Second, several factors hypothesized to be associated with HA were explored.

HWP has generally been used as the basis for determining whether a team has a HA (Courneya & Carron, 1992; Pollard, 1986; Schwartz & Barsky, 1977). For soccer, HA has been found to be about 64% (Brown et al., 1999; Courneya & Carron, 1992; Pollard, 1986). The results of this study, the first in which World Cup soccer teams were examined, were quite similar to previous findings of HWP (63%).

The consistency of the soccer HA effect is interesting in light of the fact that World Cup soccer teams do not play a balanced schedule like soccer teams studied previously (Brown et al., 1999; Courneya & Carron, 1992; Pollard, 1986). Further, World Cup teams do not typically have a home stadium. Rather, they play in several stadiums located throughout their home countries. World Cup teams also have somewhat unusual schedules of home and away competition. A World Cup team may play at home for an entire year, and not play at home at all the next year. Furthermore, each World Cup team has a unique schedule of opponents. With the differences between World Cup teams and previously studied teams, it is remarkable to note how robust the HA is for the sport of soccer.

Andersen et at. (1999), Bray (1999), and Brown et at. (1999) argued that HWP was not the ideal measure of HA, because teams' away records were not taken into account. Thus, a team with a home record of 7-5-0 and an away record of 11-0-1 would be categorized as having a HA even when no advantage of playing at home existed. Using Brown et al.'s (1999) definition of HA, most World Cup soccer teams were found to have a HA, only one team was found to have NAD, and no team had a HD. The lack of teams with NAD and HD might be expected with this sample. World Cup soccer teams consist of many of the most highly skilled soccer players in the world. Thus, it is not surprising that none of these teams had a HD.

The second component of this study was to examine several factors hypothesized to be associated with HA. It was expected that familiarity with facilities (Moore & Brylinsky, 1995); game importance (Baumeister, 1985); and travel factors (Pace & Carron, 1992) would be related to HA.

Familiarity with the Playing Facility

Researchers have suggested that the more familiar a team is with the home stadium, the greater the HA will be for that team (Moore & Brylinksky, 1995; Schwartz & Barsky, 1977; Zeller & Jurkovac, 1989). Very little statistical evidence has been provided for this hypothesis, however. Schwartz and Barsky wrote about how the manipulation of the grass levels and soil moistness of a baseball field can be used to give the home team an advantage. Moore and Brylinsky empirically tested the familiarity factors of a team that had played on the same court one year, and five different courts the next year. No differences in HA were found between the two years.

In this study of 32 World Cup Soccer teams, most games that were either lost or drawn were played on an away site. This small to medium effect size (Cohen, 1988) replicates findings of other HA researchers (Andersen et al., 1999; Brown et al., 1999; Courneya & Carron, 1992; Pollard, 1986). It seems unlikely, however, that this effect is due to players' familiarity with the stadium. As stated above, World Cup teams play their home games in many stadiums across their home country. Familiarity with the language, culture, and food may be more important contributors to this HA effect.

Game Importance

In several studies (Baumeister 1985, 1995; Baumeister & Steinhilber, 1984; Benjafield et al., 1989). teams were found to have a HD in championship series or "choke." That is, as the importance of a game increased, the HA decreased. On the other hand, several other researchers have failed to replicate the HD phenomenon in championship games (Gayton, Matthews, & Nickless, 1987; Kornspan, Lerner, Ronyane, Etzel, & Johnson, 1995; Schlenker et al., 1995a; 1995b).

For this study, championship games were won, lost, and tied just as frequently as nonchampionship games at home and at neutral sites. However, a small (Cohen, 1988) but statistically significant effect was found such that away championship games were lost more frequently than away nonchampionship games. Thus, more important away games were harder to win, which could be due to the pressure of winning an important game added to the pressure of playing at an away site.

Travel

Travel factors have been hypothesized to have a relationship with HA. Travel might adversely affect performance by causing fatigue, disrupting routines (Pace & Carron, 1994), and exposing players to climate changes (Pollard, 1986). Unfortunately, little empirical evidence has been found to support travel factors as a cause of HA. Pace and Carron (1992) did find that number of time zones crossed accounted for about 1% of the game outcome (win or lose). Courneya and Carron (1991) found that distance traveled accounted for 1.2% of explained variance in outcome.

Compared to previous researchers assessing teams that traveled in only one region of a country (i.e., the South) (Courneya & Carron, 1991) and teams that only crossed three time zones (Pace & Carron, 1992), World Cup soccer teams travel an average of 1700 km per game. In this study, the further a team traveled, the worse their performance was in terms of goals against, goals for, and goal differential. Furthermore, games that were won followed shorter trips. Although the statistical significance of the correlation between distance traveled and goals for, goals against, and goal differential may have been due to the large sample size, when games are close and all other factors are accounted for, distance traveled may have some effect on which team wins the game.

Summary

The results of this study are important for several reasons. First, highly competitive soccer teams without a balanced schedule were used in this study to demonstrate the robustness of the HA phenomenon. Soccer teams win about 64% of their games at home. Second, the results replicated the findings of Andersen et al. (1999) and Brown et al. (1999) that showed that teams with a HD tend not make it to the playoffs. Third, the small but consistent effects of travel, hypothesized to have been associated with HA, were replicated with high level athletes involved in international competitions (Courneya & Carron, 1991; Pace & Carron, 1992). Clearly, HA affects athletes of varying ability levels (Courneya & Carron, 1992). Additional research designed to assess the mechanisms by which home teams win a disproportionate number of games is warranted.

Table 1

Descriptive Statistics

Variable N Mean SD Min. Max. Skew

Distance(km) 3882 1714.51 2857.35 0.00 18674.11 2.54
Days 3882 32.22 54.88 0.00 1082.00 7.36
GA 3914 0.95 1.06 0.00 6.00 1.30
GF 3914 1.58 1.54 0.00 17.00 1.77
GD 3914 0.62 1.54 -6.00 17.00 0.80

Note. Variable names: Distance = distance from previous game; Days =
number of days from the previous game; GA = goals against; GF = goals
for; GD = goal differential.
Table 2

Descriptive Statistics for Where Played and Outcome (N = 3914)

 Win Lose Draw Total

Home
 Frequency 926 232 367 1525
 Percent 61% 15% 24% 39%
Away
 Frequency 497 458 397 1352
 Percent 37% 34% 29% 35%
Neutral
 Frequency 445 286 306 1037
 Percent 43% 28% 30% 27%
Total
 Frequency 1868 976 1070 3914
 Percent 48% 25% 27% 100%
Table 3

Pearson Product-Moment Intercorrelation Coefficients Among Variables (N
= 3882)

Variables GA GF GD Distance Days

GA -- -0.13 (**) -0.64 (**) 0.09 (**) -0.04 (*)
GF -- 0.85 (**) -0.08 (**) -0.01
GD -- -0.11 (**) 0.01
Distance -- 0.06 (**)
Days --

Note. Variable names: Distance = distance from previous game; Days =
number of days from the previous game; GA = goals against; GF = goals
for; GD = goal difference.

(*)p < .05

(**)p < .01
Table 4

Pair-wise Minimum Significant Differnces for Kruskal-Wallis One-Way
ANOVA Test Between Outcome and Distance and Outcome and Days

Group n Mean Rank Comparison AD (a) p (b)

Distance (c)

 Win(W) 1850 1772.57 W - D 235.49 < .02
 Loss(L) 968 2191.19 D - L 183.13 > .02
 Draw(D) 1064 2008.06 W - L 418.62 < .02

Days (d)

 Win(W) 1850 1991.17 W - D 93.14 < .02
 Loss(L) 968 1894.35 D - L 3.68 > .02
 Draw(D) 1064 1898.03 W - L 96.94 < .02

(a)Absolute Difference ([[alpha].sub.fw] = .10)

(b)Actual difference probability = ([[alpha].sub.fw]\k(k-1) = .02

(c)H(2) = 99.36, p<.001

(d)H(2) = 6.97, p<.05

Authors' Acknowledgements

These authors would like to thank Josh Avondoglia, Kelley Bagdasarian, Laurie BenAmi, Doug Coonrad, Janna Cunningham, Bryan Gross, Stephanie Habif, and Laura Ray for their efficient and accurate help in entering the data:

References

Andersen, M. B., Brown, T. D., Brewer, B. W., & Van Raalte, J. L. (1999). Baseball and soccer home advantage using Bray's analysis. Manuscript submitted for publication, Victoria University, Melbourne, Australia.

Baumeister, R. F. (1985). The championship choke: When a title is on the line, jubilant home crowds can root athletes on to ruin. Psychology Today. 19(4), 48-52.

Baumeister, R. F. (1995). Disputing the effects of championship pressures and home audiences. Journal of Personality and Social Psychology, 68, 644-648.

Baumeister, R. F., & Steinhilber, A. (1984). Paradoxical effects of supportive audiences on performance under pressure: The home field disadvantages in sports championships. Journal of personality and Social Psychology 47, 85-93.

Benjafield, J., Liddell, W. W., & Benjafield, I. (1989). Is there a home field disadvantage in professional sports championships? Social Behavior and Personality. 17, 45-50.

Bray, S. R. (1999). The home advantage from an individual team perspective. Journal of Applied Sport Psychology, 11, 116-125.

Brown, T. D., Andersen, M. B., Brewer, B. W., & Van Raalte, J. L. (1999, September). Soccer home advantage: New definitions and new analyses. Paper presented at the annual meeting of the Association for the Advancement of Applied Sport Psychology, Banff, Alberta, Canada.

Carron, A. V., & Hausenblas, H. A. (1998). Group dynamics in sport, (2nd ed). Morgantown, WV: Fitness Information Technology.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Courneya, K. S., & Carron, A. V. (1991). Effects of travel and length of home stand/road trip on the home advantage. Journal of Sport and Exercise Psychology, 13, 42-49.

Courneya, K. S., & Carron, A. V. (1992). The home advantage in sport competitions: A literature review. Journal of Sport and Exercise Psychology, 14, 13-27.

Edwards, J., & Archambault, D. (1989). The home-field advantage. (2nd ed.). Hillsdale, NJ: Erlbaum.

FIFA Coca-Cola World Rankings Procedures. (1999, September 14). Zurich, Switzerland: Federation Internationale de Football Association [FIFA]. Retrieved September 14, 1999 from the World Wide Web: http:www.fifa2.com/scripts/runisa.dll?M2.65816:gp::67175+/rank/ procedures1+E.

Gayton, W. F., Matthews, G. R.,& Nickless, C. J., (1987). The home field disadvantage in sports championships: Does it exist in hockey? Journal of Sport Psychology 9, 183-185.

Glamser, F. D. (1990). Contest location, player misconduct, and race: A case from English soccer. Journal of Sport Behavior, 13, 41-49.

Koppet, L. (1972, January 9). Home court: Winning edge. New York Times, pp. 1, 3.

Kornspan, A. S., Lerner, B. S., Etzel, E. F., Johnson, S., & Ronayne, J. (1995). The home disadvantage in the National Football League's conference championship games. Perceptual and Motor Skills, 80, 800-802.

Moore, J. C., & Brylinsky, J. (1995). Facility familiarity and the home advantage. Journal of Sport Behavior, 18, 302-311.

Pace, A., & Carron, A. V. (1992). Travel and the home advantage. Canadian Journal of Sport Sciences/Revue Canadienne des Sciences du Sport, 17, 60-64.

Pollard, R. (1986). Home advantage in soccer: A retrospective analysis. Journal of Sports Sciences, 4, 237-248.

Schlenker, B. R., Phillips, S. T., Boniecki, K. A., & Schlenker, D. R. (1995a). Championship pressures: Choking or triumphing in one's own territory? Journal of Personality and Social Psychology 68, 632-643.

Schlenker, B. R., Phillips, S. T., Boniecki, K. A., & Schlenker, D. R. (1995b). Where is the home choke? Journal of Personality and Social Psychology 68, 649-652.

Schwartz, B., & Barsky, S. F. (1977). The home advantage. Social Forces, 55,641-661.

Statistical Package for Social Sciences for Windows (SPSS). SPSS base 7.5 for Windows: Users guide. Chicago, IL: SPSS.

Weisstein, E. C. (1999, September 29). Great circle. Retrieved September 29, 1999 from the World Wide Web: http://www.treasure-troves.com/ math/GreatCircle.html.

Zeller, R. A., & Jurkovac, T. (1989). A dome stadium: Does it help the home team in the National Football League? Sport Place International, 3, 37-39.

Address Correspondence To: Judy L. Van Raalte, Center for Performance Enhancement and Applied Research, Department of Psychology, Springfield College, 263 Alden Street, Springfield, MA 01109. E-mail: jvanraal@spfldcol.edu.